This is the policy the product, the data model, and our human reviewers all enforce. It is public on purpose: how we handle accuracy is the whole credibility of the record.

Nothing reads as fact unless reviewed. Every case is, by default, one person’s account, posted by a user — Chirp has not verified these claims. The product never presents a grievance as established fact. Verification raises a case above that baseline; it is never assumed.

Credibility tiers (shown on every case), lowest to highest:

• Unverified · on their behalf — filed for someone else from a public source; that person hasn’t confirmed it. Lowest trust; carries a claim/remove option for the affected party.
• Unverified · user-submitted — the default for a logged-in author’s own case.
• Unverified · editorial summary — compiled by the Chirp Desk from cited public reporting; a source link is required; not independently verified or endorsed by Chirp.
• Verified affected party — the author proved they own the affected account. The ceiling for user cases.
• Right of reply present — the named company has posted a verified official response.
• Disputed — the named platform contests the case; both sides are shown and Chirp verifies neither.

What we lead with, and what we steer away from. We lead with plausible-error categories where platform mistakes are common and verifiable: wrongful/automated suspensions, billing errors, hacks and takeovers, lockouts and verification loops, glitches, dead support. We steer away from, apply higher scrutiny to, and may decline cases that are really about contested content or speech (e.g. someone seeking reinstatement after a content/conduct ban). Chirp is not an appeals court for the removed-for-cause.

Naming rules.

• Companies and platforms can be named. Criticism of a corporation, framed as a user’s experience and labeled unverified, with a right of reply available, is the point of the product. Keep claims framed as experience (“I was charged after I cancelled”), not as adjudicated fact.
• Private individuals must not be named, targeted, or exposed. Fast-track removal.
• No private/personal information. Fast-track removal.

Right of reply.Any named company can claim a verified company account and post an official response. It renders directly on the case and can move the case to “disputed.” We make this path obvious (“Represent [platform]? Respond on the record”).

Reports, takedown, and review.

• Anyone can report a case. Reports are confidential.
• Target review window: 48–72 hours for standard reports. Private-individual exposure, doxxing, and child-safety reports are fast-tracked and acted on as quickly as possible.
• A reported case stays labeled unverified while under review; reviewing a case does not endorse it.
• We maintain at least one human reviewer in the loop, a moderation queue, and the ability to hide or remove any case or response.
• Removal decisions can be appealed to [CONTACT_EMAIL].

Editorial (“Chirp Desk”) standard. Editorial cases are our own speech and are held to a journalism standard: sourced to cited public reporting, fairly summarized, right of reply offered. We summarize and link — we do not reproduce others’ full text — and we never present a grievance as established fact.

Prohibited content. Harassment campaigns or calls to brigade a person; illegal content; malware or account-cracking instructions; impersonation; spam and scams; any content that sexualizes or endangers minors. Child-safety violations override everything else and are escalated to the appropriate authorities.